Based on our work in the first cycle of the MMHCC, we propose to address two major remaining frontiers in mouse models of colorectal cancer (CRC) - opportunities afforded by a novel prevention strategy and challenges posed by liver metastasis, the major cause of mortality for individuals with CRC. Project 1 will elucidate the role of gut flora in the pathogenesis of mouse models of CRC utilizing the gnotobiotic facility at UNC. Increasing evidence strongly supports the role of the enteric flora in cancer initiation and progression. We will investigate the role of enteric flora in maintaining GI homeostasis and the role of normal flora and selected pathogens in inducing colon tumors in Smad3 null mice. We will use our germfree/gnotobiotic facility to test a standardized flora that appears to be pathology neutral, and compare with mice inoculated with Helicobacter hepaticus. We will also test various mouse colonies showing varying tumor susceptibilities and tumor distributions using a novel enteric flora profiling technique that we have developed. We propose to capitalize on this unique symbiotic relationship between enteric flora and colonic mucosa by engineering designer flora to create 'turbo-biotics', probiotic bacteria that have unique characteristics that may afford a novel and safe preventive approach to CRC. In Project 2, we will overcome a major obstacle in developing a robust model of CRC with liver metastasis - obstruction of the colonic lumen leading to bleeding and death before the tumor invades and metastasizes. We will build upon two non-obstructing mouse models of CRC that we have identified - a colony of Smad3 null mice that develop tumors only in the cecum and carcinogen-treated MOLF mice that develop sessile (flat) colonic tumors. Utilizing mouse colon cancer cell lines that can be injected into the cecum of syngeneic mice and selected for liver metastasis, we propose bold new selection schemes and a candidate approach;genes that are identified will then be introduced into our two non-obstructing models to achieve our stated goal. Employment of sophisticated imaging technologies available at VUMC will aid in this approach. In Project 3, we propose to advance progress made during the first cycle of the MMHCC. Using two classical mouse mutants, we demonstrated iterative use of EGF receptor (EGFR) signaling in intestinal neoplasia. In order to evaluate current and future therapeutic agents targeting the EGFR axis, we propose to humanize mouse EGFR. This approach will allow us to identify and validate targets of different agents that block the EGFR axis using microarray and unique proteomic capabilities coupled to strong bioinformatics support. Finally, we are continuing efforts to model pancreatic ductal carcinoma in the mouse. We have linked EGFR and Notch signaling and propose studies to elucidate roles for perturbed Notch signaling in pancreatic cancer. Although this proposal is GI-centric, the overarching goals of the bold projects with their underlying provocative hypotheses involve strategies (prevention) and processes (metastasis) central to neoplasia, and, coupled with our unique institutional capabilities, we anticipate productive interactions with other MMHCC and SPORE groups.